Saved evidence report

Fieldwork / Overview

Explore the saved analysis below. Controls filter this report; they do not rerun the analysis or retrieve source data.

Population: input · 12 evaluated source rows

Analysis: 12 rows; presence=per_row

Search coverage and limits

Search limits restrict what was tested or retained. Untested work is unknown. Display limits below only restrict this report.

missingness: 15/15 pairs evaluated

missingness: 3/3 signatures retained

dependencies: 6/6 candidate keys evaluated

paths: 15/15 pairs evaluated

paths: 6/6 features evaluated

value_patterns: 15/15 pairs evaluated

missingness

pair candidates	15
pairs evaluated	15
signatures total	3
signatures shown	3
signature omitted units	0
signature omitted rows	0
contexts total	0
contexts shown	0
entity missing key rows	0

dependencies

candidate space	6
candidates evaluated	6
dependency tests	30
contexts total	0
contexts evaluated	0
search order	determinant_size_then_input_column_order

paths

features requested	6
features evaluated	6
features omitted	none
pairs evaluated	15
pair candidates	15
paths evaluated	36
distinct alternatives	9
alternative policy	best_order_per_feature_set_collapsing_alias_substitutions
search exhausted budget	no
requested depth	2
returned depth	2

value patterns

pair candidates	15
pairs evaluated	15

Visual summary

The visual summary has its own display limit of up to 12 items per list. Browse the included evidence below.

Browse feature connections

Connections are leads to inspect, not proof of equivalence or causation.

Feature connections

attachment

mutually exclusive: attachment, callback · rows; presence=per_row · inspect evidence
exact dependency: attachment, service (attachment → service) · rows; presence=per_row · inspect evidence
strength exact
repeated support yes

callback

mutually exclusive: attachment, callback · rows; presence=per_row · inspect evidence
exact dependency: callback, channel (callback → channel) · rows; presence=per_row · inspect evidence
strength exact
repeated support yes

channel

exact dependency: callback, channel (callback → channel) · rows; presence=per_row · inspect evidence
strength exact
repeated support yes

district

equivalent value partitions: district, status · rows; presence=per_row · inspect evidence
relation equivalent_value_partitions
exact dependency: district, status (district → status) · rows; presence=per_row · inspect evidence
strength exact
repeated support yes
exact dependency: status, district (status → district) · rows; presence=per_row · inspect evidence
strength exact
repeated support yes

service

exact dependency: attachment, service (attachment → service) · rows; presence=per_row · inspect evidence
strength exact
repeated support yes

status

equivalent value partitions: district, status · rows; presence=per_row · inspect evidence
relation equivalent_value_partitions
exact dependency: district, status (district → status) · rows; presence=per_row · inspect evidence
strength exact
repeated support yes
exact dependency: status, district (status → district) · rows; presence=per_row · inspect evidence
strength exact
repeated support yes

Candidate grains

district (equivalent: status) repeated grouping

2 groups, 2 repeated; 12 determinant repeated rows; 3 exact targets, 3 with repeated support; global targets tested 5/5

columns	district
equivalent	status
role	repeated grouping
groups	2
repeated groups	2
repeated rows	12
determines	status, attachment, callback
determines with repeated support	status, attachment, callback
global targets tested	5
global targets possible	5

channel repeated grouping

3 groups, 3 repeated; 12 determinant repeated rows; 2 exact targets, 2 with repeated support; global targets tested 5/5

columns	channel
equivalent	none
role	repeated grouping
groups	3
repeated groups	3
repeated rows	12
determines	attachment, callback
determines with repeated support	attachment, callback
global targets tested	5
global targets possible	5

service repeated grouping

2 groups, 2 repeated; 12 determinant repeated rows; 2 exact targets, 2 with repeated support; global targets tested 5/5

columns	service
equivalent	none
role	repeated grouping
groups	2
repeated groups	2
repeated rows	12
determines	attachment, callback
determines with repeated support	attachment, callback
global targets tested	5
global targets possible	5

attachment constant

1 groups, 1 repeated; 4 determinant repeated rows; 1 exact targets, 1 with repeated support; global targets tested 5/5

columns	attachment
equivalent	none
role	constant
groups	1
repeated groups	1
repeated rows	4
determines	service
determines with repeated support	service
global targets tested	5
global targets possible	5

callback constant

1 groups, 1 repeated; 4 determinant repeated rows; 1 exact targets, 1 with repeated support; global targets tested 5/5

columns	callback
equivalent	none
role	constant
groups	1
repeated groups	1
repeated rows	4
determines	channel
determines with repeated support	channel
global targets tested	5
global targets possible	5

Inspect findings

12 of 28 records included. Display limit reached; increase max_findings when rendering to include more. Search only covers included records.

mutually exclusiveattachment and callback are mutually exclusive columns never populated together

Analysis: rows; presence=per_row

Finding f0 · counting unit: rows

both present	0
either present	8
both absent	4
denominator	12
presence jaccard	0
agreement	0.3333

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 5/8 saved source positions [0, 1, 2, 4, 5]

Exceptions: 0/0 saved source positions []

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f0')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

value aliasdistrict and status: equivalent value partitions equivalent encodings

Analysis: rows; presence=per_row

Finding f1 · counting unit: rows

evaluated rows	12
groups	2

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 3/12 saved source positions [0, 1, 2]

Exceptions: 0/0 saved source positions []

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f1')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

relation	equivalent_value_partitions

availabilityattachment: populated in some rows column is partially populated

Analysis: rows; presence=per_row

Finding f2 · counting unit: rows

populated	4
missing	8
denominator	12
populated fraction	0.3333

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 4/4 saved source positions [0, 1, 2, 8]

Exceptions: 5/8 saved source positions [3, 4, 5, 6, 7]

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f2')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

presence	some

availabilitycallback: populated in some rows column is partially populated

Analysis: rows; presence=per_row

Finding f3 · counting unit: rows

populated	4
missing	8
denominator	12
populated fraction	0.3333

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 4/4 saved source positions [4, 5, 6, 7]

Exceptions: 5/8 saved source positions [0, 1, 2, 3, 8]

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f3')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

presence	some

exact dependencydistrict determines status rule with repeated support

Analysis: rows; presence=per_row

Finding f4 · counting unit: rows

Exact on 12 evaluated rows; target observed on 12/12 determinant-eligible rows (observed target coverage 1); 12/12 rows in repeated groups (repeat coverage 1); repeat-only consistency 1

determinant	district
target	status
context	unavailable / not defined
exact	yes
modal accuracy	1
repair rows	0
evaluated rows	12
determinant evaluated rows	12
target observed rows	12
target coverage	1
target missing excluded rows	0
repeated rows	12
repeat coverage	1
repeat modal accuracy	1
missing excluded rows	0
evaluated groups	2
violating groups	0
affected rows	0
group violation rate	0
repeated groups	2

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 5/12 saved source positions [0, 1, 2, 3, 4]

Exceptions: 0/0 saved source positions []

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f4')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

strength	exact
repeated support	yes

exact dependencystatus determines district rule with repeated support

Analysis: rows; presence=per_row

Finding f5 · counting unit: rows

Exact on 12 evaluated rows; target observed on 12/12 determinant-eligible rows (observed target coverage 1); 12/12 rows in repeated groups (repeat coverage 1); repeat-only consistency 1

determinant	status
target	district
context	unavailable / not defined
exact	yes
modal accuracy	1
repair rows	0
evaluated rows	12
determinant evaluated rows	12
target observed rows	12
target coverage	1
target missing excluded rows	0
repeated rows	12
repeat coverage	1
repeat modal accuracy	1
missing excluded rows	0
evaluated groups	2
violating groups	0
affected rows	0
group violation rate	0
repeated groups	2

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 5/12 saved source positions [0, 1, 2, 3, 4]

Exceptions: 0/0 saved source positions []

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f5')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

strength	exact
repeated support	yes

exact dependencyattachment determines service rule with repeated support

Analysis: rows; presence=per_row

Finding f6 · counting unit: rows

Exact on 4 evaluated rows; target observed on 4/4 determinant-eligible rows (observed target coverage 1); 4/4 rows in repeated groups (repeat coverage 1); repeat-only consistency 1

determinant	attachment
target	service
context	unavailable / not defined
exact	yes
modal accuracy	1
repair rows	0
evaluated rows	4
determinant evaluated rows	4
target observed rows	4
target coverage	1
target missing excluded rows	0
repeated rows	4
repeat coverage	1
repeat modal accuracy	1
missing excluded rows	8
evaluated groups	1
violating groups	0
affected rows	0
group violation rate	0
repeated groups	1

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 4/4 saved source positions [0, 1, 2, 8]

Exceptions: 0/0 saved source positions []

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f6')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

strength	exact
repeated support	yes

exact dependencycallback determines channel rule with repeated support

Analysis: rows; presence=per_row

Finding f7 · counting unit: rows

Exact on 4 evaluated rows; target observed on 4/4 determinant-eligible rows (observed target coverage 1); 4/4 rows in repeated groups (repeat coverage 1); repeat-only consistency 1

determinant	callback
target	channel
context	unavailable / not defined
exact	yes
modal accuracy	1
repair rows	0
evaluated rows	4
determinant evaluated rows	4
target observed rows	4
target coverage	1
target missing excluded rows	0
repeated rows	4
repeat coverage	1
repeat modal accuracy	1
missing excluded rows	8
evaluated groups	1
violating groups	0
affected rows	0
group violation rate	0
repeated groups	1

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 4/4 saved source positions [4, 5, 6, 7]

Exceptions: 0/0 saved source positions []

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f7')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

strength	exact
repeated support	yes

availability signaturePresent: channel, service, district, status; absent: attachment, callback availability pattern

Analysis: rows; presence=per_row

Finding f8 · counting unit: rows

count	4
denominator	12

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 4/4 saved source positions [3, 9, 10, 11]

Exceptions: 0/0 saved source positions []

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f8')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

present	channel, service, district, status
absent	attachment, callback

availability signaturePresent: channel, service, district, status, callback; absent: attachment availability pattern

Analysis: rows; presence=per_row

Finding f9 · counting unit: rows

count	4
denominator	12

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 4/4 saved source positions [4, 5, 6, 7]

Exceptions: 0/0 saved source positions []

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f9')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

present	channel, service, district, status, callback
absent	attachment

availability signaturePresent: channel, service, district, status, attachment; absent: callback availability pattern

Analysis: rows; presence=per_row

Finding f10 · counting unit: rows

count	4
denominator	12

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 4/4 saved source positions [0, 1, 2, 8]

Exceptions: 0/0 saved source positions []

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f10')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.

present	channel, service, district, status, attachment
absent	callback

census pathchannel → attachment suggested census path

Analysis: rows; presence=per_row

Finding f11 · counting unit: rows

score

2.2

prefix counts

3, 5

prefix cost

0.2

overflow

nesting inversions

redundant steps

alias steps

target impurity sum

target impurity by depth

none

availability impurity sum

0.1667

availability impurity by depth

0.16666666666666666, 0.0

explanation

Observed prefix groups: channel: 3 → attachment: 5; Availability nonmodal fractions: 0.167, 0

reasons

kind branching
dimensions channel, attachment
prefix groups 3, 5
overflow cost 0
kind nesting
coarse to fine none
reversed edges 0
kind redundancy
no new groups none
equivalent pairs none
kind availability_separation
nonmodal fraction by depth 0.16666666666666666, 0.0

Representative source rows

Positions are zero-based offsets in the original ordered source, not dataframe index labels. Saved samples are the first matches in source order; their size is not the total support.

Examples: 3/12 saved source positions [0, 1, 2]

Exceptions: 0/0 saved source positions []

In Python, with this result named result and its identical ordered source named df:

result.inspect(df, 'f11')

Use all_matches=True to retrieve all matching rows and exceptions=True for exception rows. These operations require the original source and a Python session.